THEORIES OF FERTILITY

Robert L. Brown

Claire Norville

Abstract

Social policy planners are often faced with the difficult task of predicting future levels of fertility. A country’s fertility rate can have many political, social and economic effects. For example, the demand for social services and medical care, political views and voting patterns, the financing of pay-as-you-go social security systems are all determined by the present and past fertility rates of the population. Fertility, combined with mortality and migration, affects the age structure of the population.

Unfortunately, fertility is one of the most difficult demographic variables to predict. It can be and has been very volatile historically.

This paper reviews historic patterns of fertility in Canada and the United States. It then reviews economic theories of fertility that can be used to explain these historic patterns and could then be used to predict future levels of fertility (the review is in approximate chronological order of the theories presented).

1 I INTRODUCTION ...... 3

II THEORIES OF FERTILITY: A. RICHARD EASTERLIN...... 7 B. DIANE MACUNOVICH...... 11 C. W. BUTZ & M. WARD...... 13 D. JOHN ERMISCH ...... 17 E. A. ROMANIUC ...... 20 F. NATHAN KEYFITZ...... 21 G. A. TSUI, D. BOGUE & D. HOGAN ...... 22

III CONCLUSION ……………………………………………………………………...23

IV APPENDIX ...……………………………………………………………………….. 25

2 I INTRODUCTION

A country’s fertility rate has many short-term and long-term political, social, and economic effects. For example, the demand for social services and medical care, political views and voting patterns, the financing of pay-as-you-go social security systems are all determined by the present and past fertility rates of the population. Fertility, combined with mortality and migration, affects the age structure of the population.

Beginning in 1971, the Total Fertility Rates for the United States and Canada fell below the replacement level of 2.1. Since the 1980s, the rates for the United States appear to be increasing again and are presently approaching the 2.1 mark. However, Canada remains at levels far below replacement with a 1997 total fertility value of 1.62.

It is important that we understand what causes movements in fertility, so as to produce accurate predictions for the future. Researchers have been studying fertility rates and have produced numerous theories and projections. In this paper some of these theories are studied and their relative credence measured.

The most common measurement of fertility is the Total Fertility Rate (TFR). The TFR is the average number of children that would be born alive to a woman during her lifetime, if she was to bear children at each age according to the prevailing age-specific fertility rates. The TFR is obtained by summing the age-specific rates in a particular calendar year across all childbearing ages. Therefore, the TFR shows a cross sectional picture of fertility and consists of values from many generations of women who are at different childbearing stages in any given year.

An alternative measure of fertility is the Generational Fertility Rate (also referred to as the Longitudinal Fertility Rate). The Generational Fertility Rate (GFR) is the sum of the age-specific fertility rates over the lifetime of any birth cohort. Therefore it represents the actual number of births that a particular cohort of women experienced over their reproductive lifetime.

3 The advantage of using the GFR as a measure of fertility is that, by following a given cohort, information can be obtained on the age and intervals at which women have children. However, the major drawback of the Generational Fertility measure is that the fertility rates cannot be completely recorded until the cohort has reached the end of its reproductive years, i.e. around age 45. Therefore, the complete fertility profiles could only be obtained today for birth cohorts prior to 1955. For this reason, the discussion that follows will use Total Fertility Rates as the measure of fertility.

The graphs below (Figures 1 & 2) show the total fertility rates for Canada and the United

States. We can see that the shapes of the graphs are very similar for both countries. Of particular interest is the period from 1951 to 1966. We can see that the TFR for both countries rapidly increased, reaching peaked levels of 3.895 in Canada in 1960, and 3.682 in the United States in 1957! This generation has become known as the Baby Boom generation. We also notice the low fertility levels of the 1970’s, with the levels remaining low into the 1980’s. Canada recorded a lower trough than the United, going as low as

1.578.

Inspection of the TFR values for Quebec (see Figure 1) shows that prior to 1959 the fertility rate for Quebec was larger than that for all of Canada. However, since the Quiet

Revolution in Quebec its fertility levels have fallen below those of Canada overall.

Similarly in the United States, the TFR for Black & Other races has always been higher than that for all of the U.S. At the same time, the white population has maintained a fertility rate that is consistently slightly lower than all of the U.S. (see Figure 2).

4 Figure 3 shows a comparison of the U.S. and Canada. We can see that prior to 1966

(which marked the end of the Baby Boom period), Canada’s fertility rate was higher than that of the U.S. However, presently the Canada’s TFR is below replacement level but the

U.S. has managed to maintain a value close to the replacement rate of 2.1. Furthermore,

Figure 4 shows that the U.S. is higher than the United Kingdom, France, Japan, Germany, and Italy, (in order of descending TFR values), with Italy recording a TFR of approximately 1.2.

Clearly this has been a very volatile period for fertility rates.

What we seek to explore now are economic theories of fertility that exist that might be used to explain these past patterns of fertility, but also that could then be used to forecast future fertility rates. This review is in the approximate chronological order of the theories presented.

5 Figure 1

TFR Values for Canada

5000

4500

4000

3500

3000 CANADA (includes Quebec) 2500 QUEBEC 2000

1500

1000

50 0

0

Ye ar

TFR Values for the United States

5000

4500

4000

3500

3000 BL ACK & OTHER WHITE 2500 TOTAL

2000

1500

1000

50 0

0

Yea r

Figure 2

6 TFR for Canada and the US

4500

4000

3500

3000

2500 CANADA (includes Quebec) US (includes all races) 2000

1500

1000

50 0

0

Ye ar

Figure 3

Total Period Fertility Rates for some G7 Nations

3.50

3.00

Japan 2.50 USA 2.00 UK Germany 1.50 France 1.00 Italy

0.50

0.00 1960 1970 1980 1990 2000 2010

Ye ar

Figure 4

7 II ECONOMIC THEORIES OF FERTILITY:

A. RICHARD EASTERLIN

Richard Easterlin formulated one of the more popular theories of fertility. He believes that fertility patterns follow forty-year cycles, with large birth cohorts producing small cohorts, and vice versa. He believes that the young adult ratio in any given twenty-year period is a reflection of the birth rate of the prior twenty-year period. Easterlin says that persons born in a low birth rate period can anticipate an open and easy job market, relatively good wages, and rapid career advancement. On the contrary, persons of large birth cohorts will experience less favorable economic consequences. There are two parts to this theory: the effect birth rates have on the relative number of young adults to older adults; and the effect of relative numbers on earnings and unemployment.

The number of young adults entering the work force depends heavily on the number of births in that generation, provided that the effects of mortality and migration do not differ greatly from one generation to the next. (For Easterlin, the working population ranges from ages twenty to sixty-four, with age thirty as the dividing line between older and younger segments). Therefore, changes in the number of births between generations will correspond to changes seen twenty years later in the number of labor force entrants.

The proportion of younger men decreased in 1955-60, and then increased again until 1975. This ratio echoed the low birth rates of the 1930s and the high rates of the 1950s. The babies of the 1930s found that they were scarce compared to older workers and thus experienced little unemployment and advanced quickly in their careers. Conversely, the babies of the 1950s were in abundance relative to older workers and therefore met with unfavorable labor market conditions.

This relationship between the proportion of younger workers and unemployment can be explained by simple supply and demand arguments. When the supply of young workers is high, there will be tighter competition among them for the limited number of jobs

8 requiring younger workers. However during times of low supply, workers can choose among jobs, accepting only those offering higher wages and advancement opportunities.

Hence in the period 1945-60, when there was a scarcity of younger workers, the consequent improvement in their lifestyles resulted in a rise in marriage and childbearing. This was followed twenty years later (1970’s) by a growing abundance of younger workers and consequently a decrease in marriage and fertility. (See Appendix, Figures 6 & 7). Therefore Easterlin’s theory fits both the baby boom and baby bust periods.

The second section of Easterlin’s theory concerns the effect of relative numbers on earnings and unemployment. He hypothesized a Relative Income theory to explain the pattern. He believes that the determinants of marriage and fertility are a couple’s potential earning power; their material aspirations; and their socialization experience (e.g. religion, education and environment). The couple’s relative income is the ratio of their earnings potential to their material aspirations. Since these values are difficult to obtain, relative income is approximated by the ratio of recent income of the male to the past income of the male’s parents. In this approximation, the male’s recent income is taken as a representation of the couple’s earnings prospects. The past income of the male’s parents represents the environment in which he was raised, and therefore indicates the level of the couple’s material aspirations. Easterlin argues that as relative income increases, there is less economic pressure on the couple, and hence they are freer to marry and have children. This theory was tested against fertility rates. The results indicate that the rise in the 1940s and 1950s, and the decline in fertility from 1960 to the 1980s correspond to a similar rise and decline in the relative income approximation for couples.

Easterlin proposes that the relative income measure is, in effect, a measure of relative unemployment. This measure was also tested against fertility rates. There is an obvious similarity in the trends of relative income based on employment data and the total fertility rates of 1940 to 1980. When the outlook was good in the 1950s, marriage and childbearing increased. Conversely, when the outlook was bleak prior to World War II,

9 marriage and childbearing were deferred, and total fertility decreased. (See Appendix, Figure 5).

Looking towards the future, Easterlin predicted that from 1980 onwards, there would be a decrease in the relative number of younger workers due to the baby bust of the 1960s and 1970s. In the early 1990s he argued that the proportion of younger workers would reach the scarcity levels of the 1950s. This new change in age structure would result in a return to the patterns seen in the 1940s and 1950s. He expected there would be an increase in the income of younger workers relative to older; earlier marriages and increased childbearing within marriage; a slowdown in the increase of younger women choosing to become employed and an increase in the number of older women working; and an abatement of stagflation. Easterlin anticipated a baby boom in the 1990s comparable to that of the 1950s, with the first two decades of the 21st century resembling that of 1960 to 1980. Thereafter, the future would unfold in forty-year cycles, based on his theory.

Easterlin’s theory was published in the 1980s. Comparison with the actual TFR values experienced from the late 1980s until 1997 shows Easterlin to be incorrect in his predictions thus far. Neither Canada nor the United States has experienced the baby boom predicted, with TFR values remaining well below replacement level in Canada. It should be noted however that there was an increase in the levels of fertility for the Black and Other population of the United States during the late 1980s and early 1990s. However, this “boom” was in no way comparable to that of the 1950s. As we proceed into the 21st century, we will be able to test Easterlin’s theories for the first two decades against the levels actually observed.

10 B. DIANE MACUNOVICH

Diane Macunovich adds to the insights of Easterlin. Like Easterlin, she believes that the primary effect of cohort size (i.e. births) is the relative earnings of younger men to older men in the labor force. The decrease in relative wages is the result of the “imperfect substitutability” of the young, less experienced workers for the older, more experienced workers. However, Macunovich believes that there are two additional factors influencing relative male wages of the post-war period: the size of the military, and the size of the United States trade deficit.

The military mainly recruits young males, removing them from the labor force and consequently altering the age ratio and relative earnings of young versus older males. This effect was especially evident in the 1960s and early 1970s. Second, the level of imports tends to replace the skills of the younger, less experienced workers who are more likely to be replaced than older workers. This effect dominated in the 1980s. If both of these effects are accounted for in one’s analysis, the relative cohort size and the relative earnings of young males can be more closely approximated.

Easterlin’s theories are based on the males’ earnings and he assumes that the female plays a passive role. In her research, Macunovich recognizes the active role that young women have adopted in the face of changes in the males’ relative earnings. She believes that females also have material aspirations, but in the past, as a means of conforming to societal views, they first considered the males’ potential earnings. They used this to measure the likelihood of achieving their own personal material aspirations. But during the 1960s to 1980s some females saw that the probability of achieving their own desired standard of living based on the males’ potential earnings was increasingly unlikely. Therefore they began to project that they would need to enter the labor force and started obtaining higher levels of education. This resulted in marriage being deferred until later in life and fertility rates decreased. Thus the increase in female college entrants and increased participation in the labor force must be included when considering relative cohort size and fertility.

11 The Macunovich model measured the interaction between relative income and the female wage: an increase in the male’s relative income will cause a rise in fertility whereas an increase in female wages will produce downward pressure on fertility. In her model she suggests that during a period of high male wages the opportunity cost relative to female wages is greater than the income effect and therefore the female wage will have a negative effect on fertility. On the other hand, during a period of low male income, an increase in the female wage will result in an increase in the couple’s overall income and may therefore lead to the couple having more children. During these times the female income instead has a positive effect on fertility.

However it should be noted that this model has a one year lag between the relative income and fertility. This lag is due to the fact that couples do not immediately adjust their fertility levels whenever there is a change in their financial situation. This can be accounted for in terms of the time it takes to make the decision that they are financially prepared to have another child, the time it takes to conceive and finally the length of the pregnancy.

In forecasting future fertility rates, Macunovich supports the theory of relative cohort size as being a strong means of approximation. She points out that as the baby boom entered the work force the data supported the expectations of the theory, and similarly when the baby bust entered the labor force, the series reversed as expected. She draws reference to the fact that in 1985 the male relative wage began to increase again, and, soon after, the fertility rates of people ages 20-24 increased. Simultaneously, the entry of females aged 20-24 into the work force decreased, and the marriage rates increased and divorce rates decreased for this age group.

Maconovich concludes that the common movements in relative earnings, marriage, female labor force participation, and fertility provide strong evidence supporting the relative cohort size theory.

12 C. W. BUTZ & M. WARD

William Butz and Michael Ward challenged somewhat Easterlin’s labor supply cyclical fertility model by proposing that fertility actually moves in counter cyclical patterns to economic growth. They propose that good economic times correspond to times of high income, however, they also correspond to periods when childbirth and childrearing have a high cost in terms of the female’s time away from work. The model is based on two variables: family income and the opportunity cost of women’s time. The model hypothesizes that, all things equal, couples in higher income brackets will have more children. However, if at the same time the female is also working, the income she gives up to raise children is a significant factor. They predict that the females who earn more money will have fewer children because of the opportunity cost of childrearing being higher.

This model implies that times of economic prosperity are the most expensive times for employed women to have children. This is supported by the fact that beginning in the late 1950s, the U.S. fertility rates decreased sharply, leveling in 1970, and then decreased again until leveling again in 1974. These movements were in the opposite direction to those of the business cycle. The model suggests that the reason for this was the increase in female employment after the war. (see Appendix, Figure 8)

During the upper end of a business cycle, men’s income and consequently family income are increasing. This causes an increase in fertility rates. During recessions, the family income is lower and couples therefore decide against having children because of the high direct costs of childrearing. If we now assume that the number of jobs and the salaries for females increases steadily, then the proportion of wives that are in the work force will increase. With more women working, the business cycles will now generate cycles for both men and women’s salaries and job opportunities. Therefore, as more women are employed, times of economic prosperity no longer result in an increase in fertility. Instead, because more females are working and earning higher salaries, they will tend to delay childbirth and hence fertility will actually decrease. Thus, if enough females are

13 working, the fertility rates will begin to move in the opposite direction to the business cycle, i.e. counter cyclically.

The variables that affect the female labor force participation also affect females’ decisions to have children. Butz and Ward’s results indicate that their hypothesis fits the fertility rates of younger age groups (ages 20-24 and 25-34) and the total fertility rate of the population. However, their predictions for the group 35-39 are less accurate than for the younger groups. The reason for this is young women can decide to either delay or hasten childbirth without altering their completed fertility. This is because young women still have several more reproductive years in which to adjust their family size. However, for older women, delaying a birth usually results in a lower completed fertility because they have only a short period of fecundity remaining. Therefore older women will tend to be less sensitive, with respect to having children, to changes in family income and wages than their younger counterparts. However, Butz and Ward propose that since younger women (ages 20-24) are responsible for more than one-third of the total number of births, the overall result for the entire population is a negative relationship between female employment and fertility.

During the 1920s, TFR fell by approximately 25 percent. This decrease coincided with an increase in the disposable income per worker. Butz and Ward propose that there were two factors of interest working: females’ wages rose slowly during the 1920s and leveled off in 1935; at the same time, in 1934, the disposable income per worker increased sharply. Historical data shows that the fertility rate decreased during the 1920s and then leveled off in 1935. Butz and Ward believe that the increase in female wages was (partially) responsible for the decrease in fertility in the 1920s. But then around 1935, the combination of rising disposable income per worker and decreasing female wages caused the fertility rate to level off.

Butz and Ward next considered the data of the 1950s to 1970s. In the 1950s, fertility rates tended to follow the business cycle because the proportion of females in the work force was not large enough to have an opposing effect on the relationship. The 1960s was a

14 time of economic expansion and was therefore accompanied by increased wages and job opportunities. As more women became employed, pregnancy was delayed. Then in 1970, the economic recession caused women’s salaries to decrease and there was a reduced number of job opportunities. This recession caused fertility to increase since the cost of childrearing was low with regards to the wife’s foregone income. After the recession, the number of women in the workforce increased dramatically as the wages and job opportunities soared. This increased employment meant more women were choosing to work instead of having children. This increase in fertility followed by a decrease was repeated again during and after the 1974 recession.

In summary, over the period from the late 1940s to the 1970s, there were three major intervals over which fertility rapidly declined: 1947-50, 1957-69, and 1970-73. During these periods employment and wages increased. Similarly, during the periods 1950-57, 1969-70, and 1973-74, fertility either increased or very slightly decreased. During these periods, wages were either declining or constant, with 1969-70 and 1973-74 exhibiting constant or slowly increasing employment ratios.

This experience supports Butz and Ward’s hypothesis that fertility rates are positively related to family income, and negatively associated with women’s employment and wages. The correlation between women’s wages and fertility is stronger, the larger the proportion of women employed.

The model was used to predict fertility for 1956 to 1975. It predicted a higher and larger fertility peak than the one actually experienced in the late 1950s. It was successful in predicting the rapid decline in fertility during the economic expansion of the 1960s, and the movements during the 1970 and 1974 recessions.

The Butz and Ward model explains both pro-cyclical and counter-cyclical fertility movements. Looking to the future, their model implies that, provided female wages and job opportunities continue to increase, the fertility rate will decline until a bottom is reached. They predict that counter-cyclical movements due to couples having children

15 during slower economic times will interrupt the decrease. This is the opposite of Easterlin’s prediction, which states that fertility rates will move on an upward path. However, according to Butz and Ward, the decrease in rates will result in a top-heavy age structure by 2020.

16 D. JOHN ERMISCH

John Ermisch, like Butz and Ward, debates the veracity of Easterlin’s theory. He believes the increasing demand for female workers, and not movements in relative income, is the main explanation of fertility movements. Ermisch’s hypothesis, like that of Butz and Ward, proposes that the major influence on a couple’s family size is whether one or both partners work.

Usually males remain in the work force throughout their lives but female lifetime labor force participation is more variable. Some females opt to work throughout their life, breaking for only short time intervals during which they bear children. Other women instead leave the labor force upon the birth of the first child and then return after the last child has entered school, or sometimes they never return. Therefore women can be divided into two very broad groups: “workers” and “non-workers”. The labor force participation probability of “workers” tends to be close to one, whereas that of “non- workers” tends to be close to zero. The factors which can permanently influence a female’s decision to work include: her husband’s expected future earnings, her earning potential, and changes in preferences in the household. If these “permanent influences” cause the female to enter the work force, she is more likely to remain there with brief breaks for childrearing.

Childrearing is one of the most time consuming of a couple’s “home-produced commodities”, and therefore the opportunity cost of having children is high. Therefore increases in the value of time will increase the cost of having children, and decrease the demand for children. In families where the male is the only bread earner, most of the child-care will be the responsibility of the female. Increases in the husband’s salary will increase the family income. The result is an increase in the demand for children.

In two earner families, the husband tends to spend more time participating in child-care at the expense of his leisure time. The cost of both spouses’ time is affected by changes in their individual future earnings. Since they are both contributing to the childrearing, an

17 increase in their expected earnings will increase the cost of having children, and will have a net effect of a decrease in their demand for children. A child would demand more of the couples’ time and lower the family’s present income due to the loss of the wife’s earnings during the early years. Therefore, in double earner families, an increase in wages will raise the value of leisure and work time, thus increasing the opportunity cost of children and reducing the couple’s desired family size.

Combining these two cases of single and double earner families implies that the effect of salary changes on the average family size depends on the number of couples in each of these groups (i.e. one or two-earner families). When there are more female “workers” than “non-workers”, fertility tends to decrease during times of economic growth. However, if there is a larger proportion of “non-workers”, fertility rates will increase during these times of economic growth.

In his research, Ermisch also discovered that as more females choose to work most of their lives, the average age at first birth increases and the intervals between births decrease. He argues that women employed in managerial and professional positions, (representing most of the female “workers” population), tend to wait longer periods between marriage and the birth of their first child than other women occupying intermediate non-manual and clerical positions. Similarly, women in the latter group tend to wait longer than females in manual occupations. (This can be seen for example in the fertility observed in Australia. According to the Australian National University, the fertility is now concentrated in ages 25-34. The average age at first birth for females has increased from age 23.7 in 1971 to age 28.5 in 1995. The median age of mothers giving birth in a year was 26.5 in 1971 whereas it was recorded as 28.9 in 1995.)

Ermisch also recognizes the inter-generational influences on females’ labor force participation. He believes that daughters of working mothers tend to be more likely to work during their childbearing years. Therefore there is an indirect effect on a woman’s fertility from her mother being employed, by changing the number of women opting to become “workers” versus “non-workers”.

18 Looking to the future, Ermisch believes that changing economic conditions will cause couples to adjust their desired family size. During economic growth the proportion of two earner families with the minimum desired family size would increase. Simultaneously, economic growth has the effect of increasing the desired family size of single earner families. In sum, economic growth will continue to exert a depressing effect on fertility of two-earner couples. Thus the average family size will be reduced until we reach a lower asymptote. This means that if all couples were double earners then the fluctuations in fertility would move in the opposite direction to that of the business cycle, as proposed in the Butz and Ward model. This is counter-balanced by the cyclical nature of fertility if there is a larger number of one-earner families. Thus, this theory leads to a prediction of the gradual flattening in fertility rates, with no sustained increase in the future.

19 E. A. ROMANIUC

Romaniuc recognizes that Canada is presently in its second low fertility cycle. The first cycle occurred between the two World Wars and the second one began at the beginning of the 1960s and continues through to today. However, these two cycles differ significantly in magnitude. The difference in birth rates among the regions has narrowed compared to their rates prior to World War II. Additionally, there has been a decrease in the effects of factors such as ethnicity, language, and religion. However, there still exist differences due to labor force participation and education.

Overall, fertility rates at the ends of childbearing years, i.e. ages 20-24 (lower end) and 40-44 (upper end), have decreased and those at ages 30-34 have shown a slight increase. This is the result of women waiting longer to have their first child and giving birth at longer time intervals.

Romaniuc conducted surveys in the United States and Quebec, which indicate that there is a significant decrease in the number of children that young couples expect to have. There are several possible reasons for this decline, such as changes in women’s roles, and changes in the status of children. There appears to be a definite increase in the number of young couples intentionally choosing childlessness. This trend is currently assisted by increased use of contraception, family planning and social equality in sex roles.

A further factor exerting a downward pressure on fertility is the decrease in marriages and the rise in divorce rates. Romaniuc discovered that there is a significant reduction in the marriage rate, particularly prior to age 35. The divorce rate of 1981 was six times that of 1960-1962. These effects outweigh those of increased out-of-wedlock births.

In the 1970s, Canada experienced a slowdown in economic growth, an increase in unemployment and inflation. These economic fluctuations may not be the main reason for the decline in fertility rates, but they may be exerting conditioning influences. Since the 1960s, the baby-boomers have had to deal with the large size of their generation and

20 also with the problem of entering the labor force during a period of rising inflation and economic slowdown. This dual problem is referred to as the “economic squeeze” by Openheimer, and arises when lifestyle aspirations exceed the couple’s purchasing power (see Easterlin). In response to this “economic squeeze” marriages and childbearing are deferred, and there is an increase in preference for small family size.

F. NATHAN KEYFITZ

N. Keyfitz, a professor of Demography, also studied the effects of factors such as abortions and contraception on the level of births observed.

Keyfitz proposes a model which implies that an abortion does not prevent a birth in the population as basic logical thinking suggests. Instead it could actually have the effect of reducing the time until the female’s next pregnancy. The act of abortion could on average add a fraction of a birth. The mathematical proof of this is beyond the scope of this paper.

Keyfitz research indicates that females using contraception tend to have all their children within a few years during their fecundity period. As the level of efficiency of the contraceptive increases, the more concentrated the childbirth period becomes. Therefore instead of the mother being out of the work force on several occasions for the rearing of each child, she instead leaves on one occasion. Keyfitz believes that by encouraging couples to have their children in a clump instead of spread over 25 years, birth control could actually be increasing the rate of growth of the population. His research discovered that older women seem to have been attracted to birth control before younger women. This results in a sharper decrease in childbirths in the older reproductive years than in the younger.

21 G. A. TSUI, D. BOGUE & D. HOGAN

Tsui, Bogue and Hogan believe that the decrease in the fertility rates of the 1970s may be due to changes in fertility within marriages (nuptial fertility); changes in extramarital fertility; and changes in the number of marriages. They suggest that variables such as the couple’s demographics (e.g. age of each partner, marital status, duration of marriage, number of living children); each partner’s background (i.e. race, ethnicity, language, urban-rural origin); the female’s education level; and each person’s occupation should be included and controlled in any fertility analysis.

Movements in fertility are strongly influenced by the “childbearing attitude climate” of the couple. This attitude determines the desired family size, the sex preference for children, feelings towards use of contraception and results in questions such as “Was the last child born wanted or unwanted?”.

Desired family size is self-explanatory. Sex preference for children refers to a couple’s need to have a certain number of sons (daughters). Therefore their childbirths are less likely to be determined by their desired family size and more by the total number of sons (daughters), or combination thereof, which they aspire to have. The couple’s attitude towards the use of contraception serves to separate the population into two distinguishable groups governed by their positive attitude (known as “pro-birth- planning”) or negative attitude (known as “pronatalist”) towards contraception. Therefore, the overall fertility movements seen in the population are affected by the proportion of people in each group. Finally the issue of unwanted versus wanted last pregnancy determines whether the couple’s last encounter was one that left a feeling of pleasure or one of disappointment and other negative emotions. These feelings in turn affect their attitude towards possible future pregnancies.

Tsui, Bogue and Hogan suggest that there is an inverse relationship between the couple’s level of contraceptive knowledge and their desired family size. Their research indicates that couples with small desired family sizes tend to be very knowledgeable on fertility- regulating practices. This in turn allows them to attain and not surpass their family size

22 aspirations. However, they recognize that there isn’t an obvious direction of causation, (i.e. Does the possession of information on contraceptives reduce family size aspirations or does the preference for a small family cause couples to acquire contraceptive information?). The answer to this question is still unknown.

23 III CONCLUSION

Easterlin’s theories explain the fertility patterns seen in the early years (i.e. pre-1980) when the effect of female employment was negligible. Easterlin assumes that the female enters the labor force only if the male’s income is insufficient, otherwise she devotes herself entirely to the running of the household. His model therefore fails to explain the trends seen in the 1980s and 1990s when more females were employed. According to the demographic ratios of 1921 to 1995, Easterlin’s theory does not hold for the entire period, instead it can only be confirmed for the period 1940 to 1980.

Pampel and Peters (1995) suggested some reasons for the inaccuracy of Easterlin’s theory post-1980. These include the increasing effect of the business cycle on the labor market demand; a recent increase in the number of immigrants causing greater labor force competition; increasing female labor force participation; and long term effects on the job market of the exceptional size of the baby boom. For these reasons, the author would not recommend the Easterlin cycle to be used for future fertility predictions.

However it should be noted that although real wages have increased by a factor of 3.6 over the period 1921-1991, couples are still having fewer children than those of the previous generations. This may actually serve to support Easterlin’s hypothesis that the additional income is used to satisfy the couple’s material needs, which tend to be higher than in previous generations, rather than being used to support another child. If this is true then Easterlin would have correctly explained the recent trend of limiting the number of children per household as a means of maintaining a high standard of living.

The Macunovich model is based heavily on Easterlin’s male income hypothesis but it also includes a variable for changes in the level of the female’s wages. She then goes even further to determine if an interaction exists between these two conflicting variables. Her model has produced highly accurate results for the period 1969 to 1993. It even accurately accounted for the small “boomlet” of the 1980s. By combining the two

24 theories and overcoming their limitations Macunovich has produced an accurate means of predicting future fertility levels..

Butz and Ward propose a theory of counter cyclical patterns and take into account the opportunity cost of childrearing. Their theory manages to explain the trends in later years (post-1980) when more females worked. However, it does not adequately deal with the early time periods, i.e. prior to 1940, when females generally did not work. However, if we only require predictions of future fertility levels this theory would probably produce fairly accurate estimates.

Similarly, the theories of John Ermisch could produce fairly accurate estimates. However, this theory tends to be slightly complicated by taking into account a large number of variables, e.g. number of one-earner relative to two-earner families; whether or not the female works; the present state of the economy etc. Even though Ermisch is correct in stating that each variable will affect the fertility level seen, this author believes that when making predictions this leads to several places where errors could be made with respect to assumptions about future states, thus potentially decreasing the accuracy of the estimates.

Romaniuc, Keyfitz, Tsui, Bogue and Hogan are all correct in proposing that the fertility levels depend on such variables as age of first birth, contraceptives and family planning. However, this author believes that these factors on their own are not sufficient to explain fertility trends. Instead these factors should be taken into account when using one of the other theories above, to produce more accurate predictions.

In summary, the movements of fertility will continue to cause debate. There is no single factor or theory to explain its trends. The combined effect of several factors was responsible for the baby boom: a thriving economy; fewer people entering the labor market; modest material goals set by couples who grew up during the Depression in the 1930s; and men’s relative salaries were more comfortable than in the past. A number of important variables have been identified over the years, but the list is by no means exhausted. Fertility will continue to fascinate some of the world’s best scholars.

25 APPENDIX Total Fertility Rates per 1000 women for Canada: CANADA CANADA CANADA (includes (includes (includes YEAR Quebec) QUEBEC YEAR Quebec) QUEBEC YEAR Quebec) QUEBEC 1921 3536 1951 3503 3775 1980 1676 1616 1922 3402 1951 3503 3775 1981 1648 1569 1923 3234 1952 3641 3861 1982 1637 1483 1924 3221 1953 3721 3877 1983 1625 1435 1925 3132 1954 3828 3944 1984 1630 1427 1926 3357 4307 1955 3831 3904 1985 1613 1400 1927 3319 4266 1956 3858 3904 1986 1594 1375 1928 3294 4195 1957 3925 4001 1987 1578 1366 1929 3217 4010 1958 3880 3938 1988 1607 1428 1930 3282 4059 1959 3935 3928 1989 1660 1528 1931 3200 4001 1960 3895 3764 1990 1712 1638 1932 3084 3804 1961 3840 3700 1991 1699 1650 1933 2864 3502 1962 3756 3578 1992 1693 1650 1934 2803 3441 1963 3669 3473 1993 1658 1611 1935 2755 3369 1964 3502 3333 1994 1662 1612 1936 2696 3364 1965 3145 2996 1995 1639 1579 1937 2646 3268 1966 2812 2646 1996 1552 1522 1938 2701 3261 1967 2597 2367 1997 1624 1601 1939 2654 3211 1968 2453 2180 1940 2766 3287 1969 2405 2100 1941 2832 3389 1970 2331 1974 1942 2964 3529 1971 2187 1878 1943 3041 3571 1972 2024 1727 1944 3010 3643 1973 1931 1683 1945 3018 3666 1974 1875 1657 1946 3374 3832 1975 1814 1659 1947 3595 3896 1976 1765 1680 1948 3441 3805 1977 1746 1678 1949 3456 3797 1978 1698 1631 1950 3455 3812 1979 1700 1670

26 Total Fertility Rates per 1000 women for the United States: BLACK BLACK BLACK & & & YEAR OTHER WHITE TOTAL YEAR OTHER WHITE TOTAL YEAR OTHER WHITE TOTAL 1917 3451 3312 3333 1946 2937 2843 2858 1975 2276 1686 1774 1918 3475 3288 3312 1947 3247 3167 3181 1976 2223 1652 1738 1919 3344 3025 3068 1948 3402 2968 3026 1977 2279 1703 1790 1920 3556 3219 3263 1949 3509 2964 3036 1978 2265 1668 1760 1921 3615 3282 3326 1950 3579 2945 3028 1979 2310 1716 1808 1922 3354 3072 3109 1951 3718 3120 3199 1980 2177 1773 1840 1923 3332 3063 3101 1952 3760 3213 3287 1981 2118 1748 1812 1924 3454 3069 3121 1953 3880 3269 3349 1982 2107 1767 1828 1925 3417 2949 3012 1954 4053 3372 3461 1983 2066 1741 1799 1926 3324 2839 2901 1955 4126 3405 3498 1984 2071 1749 1807 1927 3101 2783 2824 1956 4286 3504 3605 1985 2109 1787 1844 1928 2829 2632 2660 1957 4347 3582 3682 1986 2136 1776 1838 1929 2716 2506 2532 1958 4285 3532 3629 1987 2198 1805 1872 1930 2734 2506 2533 1959 4321 3537 3638 1988 2298 1857 1934 1931 2641 2369 2402 1960 4238 3510 3606 1989 2433 1931 2014 1932 2668 2271 2319 1961 4234 3464 3564 1990 2480 2003 2081 1933 2523 2126 2172 1962 4108 3320 3423 1991 2480 1996 2073 1934 2605 2181 2232 1963 4000 3194 3298 1992 2442 1994 2065 1935 2548 2141 2189 1964 3900 3061 3171 1993 2385 1982 2046 1936 2483 2101 2146 1965 3808 2783 2913 1994 2300 1985 2036 1937 2563 2121 2173 1966 3532 2603 2721 1995 2175 1989 2019 1938 2584 2175 2222 1967 3299 2447 2558 1996 2149 2019 2040 1939 2564 2120 2172 1968 3108 2366 2464 1997 2158 2017 2040 1940 2618 2177 2229 1969 3061 2360 2456 1941 2690 2281 2332 1970 3067 2385 2480 1942 2744 2526 2555 1971 2920 2161 2267 1943 2837 2611 2640 1972 2628 1907 2010 1944 2787 2452 2495 1973 2444 1783 1879 1945 2733 2375 2422 1974 2339 1749 1835

27 Total Period Fertility Rates:

COUNTRY 1960 1970 1980 1990 2000 2010 Japan 2.02 2.07 1.76 1.49 1.47 1.61 USA 3.31 2.02 1.82 2.05 1.93 1.90 UK 2.81 2.04 1.80 1.78 1.72 1.83 Germany 2.49 1.64 1.46 1.30 1.33 1.44 France 2.85 2.31 1.87 1.72 1.75 1.89 Italy 2.55 2.28 1.55 1.28 1.22 1.33

28 29 30 31